Panel Data Reduces Bias in Entry Models

Abstract

Entry Models such as Bresnahan and Reiss(91)[4] can under estimate the effect of competition. If the profitability of marketsis mismeasured, this introduces an positive correlation between unoberserved profitability and the number of firms in a market. Using data on entry and exit patterns in the Ready-Mix Concrete Industry from 1976-1999, I show that using fixed effectsin a Bresnahan-Reiss entry model reduces the coefficient on demand by 50% and increases the coefficient on competition by 100% compared to the no fixed effect benchmark.